NVIDIA
By downloading or using the Model, you agree to the terms of the licenses linked to this page.
Copied!
Model
NVIDIA
Nemotron-3.5-Super-VL-120B-A12BNVIDIA Nemotron 3.5 Super VL 120B-A12B vision-language model weights (BF16, NVFP4) used by the NIM.
By downloading or using the Model, you agree to the terms of the licenses linked to this page.
Copied!
| Field | Response |
|---|---|
| Participation considerations from adversely impacted groups protected classes in model design and testing: | None |
| Bias Metric (If Measured): | BBQ Accuracy Scores in Ambiguous Contexts |
| Which characteristic (feature) show(s) the greatest difference in performance?: | The model shows high variance in the characteristics when it is used with a high temperature. |
| Measures taken to mitigate against unwanted bias: | Supervised Fine-Tuning (SFT) and Reinforcement Learning from Human Feedback (RLHF) employed to calibrate the model’s reasoning capabilities to maintain logical consistency and appropriate complexity when interacting with or interpreting data from diverse age demographics. |
| If using internal data, description of methods implemented in data acquisition or processing, if any, to address the prevalence of identifiable biases in the training, testing, and validation data: | The training datasets contain a large amount of synthetic data generated by LLMs. We manually curated prompts. |
| Tools used to assess statistical imbalances and highlight patterns that may introduce bias into AI models: | BBQ |
| Tools used to assess statistical imbalances and highlight patterns that may introduce bias into AI models: | These datasets, such as web-scraped finance reasoning data derived from SEC EDGAR filings, science and math problem datasets, OpenResearcher/source-document datasets, Common Crawl, CC-News, Wikimedia, long-context document datasets, and image datasets including BLIP3, PixMo, COCO-derived collections, and document, chart and page-image datasets, do not collectively or exhaustively represent all demographic groups (and proportionally therein). |
| For instance, these datasets do not contain explicit mentions of demographic classes such as age, gender, or ethnicity in approximately 97% to 99.9% of finance reasoning samples and in over 85% of samples across the broader assessed text datasets, and automated person-presence screening did not detect a person in approximately 86% of the images analyzed across 40 image datasets (and in over 98% of images in 18 document-, chart- and filing-page-focused datasets). In the subset where such terms or people are present, these datasets contain representational skews; for example, in text, ethnicity mentions are often dominated by Middle Eastern contexts (found in finance documents) or "White," "Two or more," and "Black or African American" as the most frequent ethnic identifiers, while references categorized as male-only significantly outnumber those categorized as female-only. In image datasets, where person-present signals were identified, these datasets contain uneven representation splits across the measured visual taxonomies: apparent age was predominantly adult, followed by elder; gender presentation skewed male; and automated apparent-ethnicity candidates were led by White and Asian. To mitigate these imbalances, we recommend considering evaluation techniques such as bias audits, task-specific fairness evaluation, red-teaming, fine-tuning with demographically balanced datasets, and mitigation strategies such as counterfactual data augmentation to align with the desired model behavior. This evaluation used a 1,000-3,000-sample subset per dataset, identified as the optimal threshold for maximizing embedder accuracy, and includes outputs from uncalibrated embedders; as such, certain limitations may exist in the reliability | |
| Unwanted Bias Testing: | Constrained to English-language inputs. Multi-lingual parity is not currently claimed or guaranteed. |